'BootCD' 카테고리의 다른 글

Boot CD Link  (0) 2009.07.01

http://support.citrix.com/article/CTX121875

Physical Volume Enabling an HVM Linux Virtual Machine imported from VMware

http://www.youtube.com/watch?v=2EaixwuM6Vw
앞서 설명드린 HPC Server 2008 Job Resources 에서는 Job Scheduler를 사용하는 경우였습니다.
그런데,, 아래 그림처럼 다른 툴을 이용해서 job submit 을 하는 경우에는 적용하기가 어렵습니다.


그래서 약간 우회적인 방법으로 노드당 1개의 core만 사용할 수 있는 방법을 알려드립니다.


임의의 job을 submit 해서 각 노드의 core를 점유한 후에 Fluent Launcher 를 실행하는 방법입니다.

예를 들어 hostnode, node001, node002, node003, node004 5대의 서버가 있고,
각 서버당 Quad-Core CPU 1개씩 장착되어 있다면,, 총 사용 가능한 core는 20개 입니다.
5개의 각 노드당 1개의 core 씩만 사용해서 job을 수행하고 싶다면,,
  1. HPCS Job Scheduler 를 실행해서 New Job
  2. Run job until cancelled or run time expires 선택
  3. Core 단위로 MIN=3, MAX=3 선택
  4. Resource Selection 서 hostnode(첫번째노드) 선택
  5. job submit
  6. node001, node002, node003, node004 에 대해서도 동일하게 진행
  7. Fluent Launcher 를 실행해서 5개의 프로세스 요청
위와 같이 진행하면, 이미 모든 노드에서 3개씩 core를 잡고 있기 때문에,
신규로 Job 이 들어갈 경우 각 노드 당 1개의 core만 사용하게 됩니다.




'Scheduler' 카테고리의 다른 글

[SGE] Rocks 5.3 + mpich2  (0) 2010.08.07
[PBS] Torque+Maui 스케쥴러 설치  (0) 2009.06.21
[SGE] Parametric Job + Job Dependency  (0) 2009.06.21
HPC Server 2008 Job 생성 메뉴에 보면,, Job resources 라는 항목이 있습니다.
Job을 Core 마다 할당하거나 Node 마다 할당할 수 있습니다.

WCCS 를 사용하는 경우에는 run1perNode.vbs 와 같은 스크립트의 도움을 받아야
(다음 링크 참조 :
http://www.alwaysu012.net/18Node 당 1개의 프로세스를 할당할 수 있었지만,,
HPC Server 2008 의 경우에는 Job resources 에서 선택해서 간단히 해결할 수 있습니다.










ISO 이미지 사용
xe sr-create name-label=<name> type=iso \
device-config:location=<where iso file exist> \
device-config:legacy_mode=true \
content-type=iso

EX) /var/opt/xen/iso_import/
xe sr-create name-label=windows_iso type=iso \
device-config:location=/var/opt/xen/iso_import/ \
device-config:legacy_mode=true content-type=iso

[root@xen12 ~]# xe sr-list name-label=windows_iso
uuid ( RO)                : 7bdf8f9c-ba67-eeba-ff59-ec590f8f3692
          name-label ( RW): iso-file
    name-description ( RW):
                host ( RO): xen12
                type ( RO): iso
        content-type ( RO): iso

sftp 와 같은 툴을 이용해서 ISO 파일을 /var/opt/xen/iso_import 디렉토리에 업로드한 뒤
스캔해서 업로드된 ISO 파일을 인식시킨다.

[root@xen12 ~]# xe sr-scan uuid=7bdf8f9c-ba67-eeba-ff59-ec590f8f3692

정말 급변하는 시대이다.
잠시 한눈 팔면,, 변화에 뒤쳐지기 마련이다..
최근에 HPC Cluster 구축 요청을 받으면서,,
고객은 이미 한발 앞서가고 있는데,, 서비스를 제공한다는 사람이 관련 내용을 전혀 모르고 있다니 !!
급 실망.. ~~~~
하지만 늦었다고 생갈할 때가 가장 빠르다는 격언을 되새기며,, 최근의 Cluster 툴을 살펴보자.

1. http://www.rocksclusters.org
 - 가장 애용하는 클러스터 툴이다. 이 툴을 접해본지도 벌써 5년이 다되간다.
   무엇보다 구성하기가 쉽다. 그동안 구축한 reference 도 상당히 많다.
   앞으로도 가장 많이 사용하게될 툴....

2. http://www.clustercorp.com (Rocks+)
 - Rocks Cluster 는 OpenSource 이다. 이 패키지 안에 포함되어 있는 툴들도 모두 OpenSource 이다.
   하지만,, 클러스터를 구성하다 보면 상용툴을 설치해야 하는 경우가 종종 있다.
   예를 들면, Intel Compiler, TotalView, PBS Pro 등 상용 Scheduler 등등..
   ClusterCorp는 OpenSource Rocks 를 기반으로 하며, 상용 툴들에 대한 Roll CD를 제공해 준다.
   Roll CD를 다운로드 받으려면 반드시 등록된 ID가 있어야 하며,,
   구매 가격은 $149/node 부터 시작한다.

3. http://rocks.stanford.edu/doku.php
  - 상용 툴을 사용하기 위해 반드시 반드시 Rocks+ 를 구매해야 하는 것은 아니다.
    위 싸이트에 접속하면 Intel, PGI, openIB, MPI Library 등을 구할 수 있다.

4. http://www.redhat.com/hpc
   http://www.platform.com/cluster-computing/cluster-management
   http://www.hpccommunity.org/
  - 이 툴은 정말 최근에야 소식을 접했다
    예전에 Platform 에서는 Rocks를 기반으로 Open Cluster Stack (OCS) 을 제공했는데,
    최근에 PCM (Platform Cluster Management)  으로 그 툴이 변경되었다.
    KUSU 라는 프로젝트를 기반으로 하고 있다는데,,
    알아봐야 할 내용이 많은 부분이다.
  - RedHat HPC 를 구매하면 yum을 통해서 PCM을 구성할 수 있다고 하는데,,
    현재 Evaluation 중인데, 완전 까막눈이라 갈 길이 멀다.

5. http://www.penguincomputing.com/software/scyld_clusterware
   - Scyld ClusterWare 라는 클러스터 구축 툴이다.
     얼마전 S대에 펭귄클러스터가 납품되었는데, diskless 구성이다.
     사용해 본 분이 그 어떤 툴보다 사용이 편리하다고 하더라...
     Evaluation 을 해봤는데,, 구축하기도 쉽다.
     좀 더 테스트가 필요한 툴이다.

6. http://svn.oscar.openclustergroup.org/trac/oscar/wiki
   - OSCAR
   - 한번도 시도해 보지 않은 툴이다.
      현재도 계속 업데이트되고 있는데,, 언젠가 한번 테스트 해봐야지.

7. http://xcat.sourceforge.net/
  - xCAT Extreme Cloud Administration Toolkit
   2005년 이후에 설치해 볼 기회가 없었다.
   IBM에서 OpenSource 로 넘어오면서, 구조 자체가 바뀌었다고 하던데,,,
   언젠가는 테스트해 볼 기회가 오겠지..
현재 다운로드 가능한 버전은 5.3 Tacos 입니다.
Rocks 5.3 (Rolled Tacos) for Linux

Rocks 홈페이지에 방문하면 오른쪽 하단에 Google Calendar 가 있습니다.
여기에 보면 9월 14일 쯤에 5.4 버전이 출시된다는 군요..

Rocks-A-Palooza 에 가보면,, 5.4에 추가되는 기능이 나옵니다.
http://www.rocksclusters.org/rocksapalooza/2010/whats-in-5-4.pdf

NIC Bonding 기능이 눈에 띠네요..
요즘엔 서버 또는 PC에 장착된 CD-ROM 이 IDE/SATA 방식이 아니라,
USB 방식으로 연결되어 있는 경우가 있습니다.
이럴 경우 초기에 Kernel 이 올라온 후 CD-ROM 을 못찾는다는 메시지와 함께,,
더이상 설치를 진행할 수 없게 되죠..

The Rocks CD was not found in any of your CDROM drives.  Please insert the Rocks CD and press OK to retry.

이럴 때 아래 방법을 사용해 보세요..

https://wiki.rocksclusters.org/wiki/index.php/Tips_and_tricks

 

Q: How do I use a USB attached DVD drive to install the frontend?

A. Append driverload=usb-storage to the build command when you first boot the frontend.


# build driverload=usb-storage

6월, 11월은 TOP500 순위가 발표되는 달입니다.
때를 맞춰 HPC에 관련된 회사들은 신제품을 대거 출시하죠..
세미나에서 제품 발표도 하고,,
WCCS -> HPC Server 2008 -> HPC Server 2008 R2 의 출시 시기를 보면,,
6월, 11월에 맞춰져 있는 것도 이 때문입니다.

이번에도 어김없이,, HPC Server 2008 R2 Beta1을 출시했네요..
11월 초에 발표되었습니다.
내부적으로는 Version 3라고 부른다고 하네요..
Windows Compute Cluster Server 2003 = V1
Windows HPC Server 2008 = V2
Windows HPC Server 2008 R2 = V3

다운로드 받으려면 connect.microsoft.com 에 로그인할 수 있어야 하구요..
싸이트에 접속하면,, Excel 관련 패키지가 보입니다.
cnet news의 제목을 보면,,
"Microsoft testing Excel for supercomputers"
http://www.microsoft.com/presspass/press/2009/nov09/11-16SC09PR.mspx

V1에서 Excel 과 CCS를 연동하려면, UDF를 작성해야 하고,,
번거로움이 많았는데,, 이제는 어떻게 해결했다는 건지,, 궁금하네요..

설치해 봐야 하는데,, 요즘 너무 게을러 져서....

ssh와 sftp는 분명히 같은 포트(Port 22)를 사용하는데,,
ssh 는 정상적으로 접속이 되고,, sftp는 접속이 안된다??
정말 황당 사건입니다.

유사한 사례를 살펴보니,,
Shell Startup Scripts 에서 echo 에 의한 ouput이 있는 경우
이런 문제가 발생할 수 있다고 하네요..

문제가 발생하는 경우,,
1. ssh 접속했을 때 예전과 다른 문자가 찍히는지 확인
2. 문자가 찍힐 경우 아래와 같은 startup scripts 를 열어서,,
   echo 에 의한 출력이 있는지 확인해 보세요..

$HOME/.bashrc
$HOME/.profile
/etc/bashrc
등등

3. 출력되는 부분이 있을 경우,, 지워보세요..

'Linux' 카테고리의 다른 글

Linux Backup & Restore  (0) 2010.08.02
[UDA] OS 자동설치 툴  (0) 2010.08.02
[TIP] Shell Script  (0) 2009.07.01
Play Station 3 의 굴욕  (0) 2009.06.30
Fedora CORE 설치 중 hpet 메시지에서 멈출 때..  (1) 2009.06.25
Rocks 5.3에서 samba 서버를 사용할 경우
iptables Firewall 때문에,, 외부에서는 접속이 안됩니다.
아래와 같은 명령으로 samba 관련 포트를 열어야 사용할 수 있습니다.

/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto tcp --dport 137 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto tcp --dport 138 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto tcp --dport 139 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto tcp --dport 445 --jump ACCEPT

/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto udp --dport 137 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto udp --dport 138 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto udp --dport 139 --jump ACCEPT
/sbin/iptables --insert INPUT --source your.ip.networks.0/255.255.255.0  --in-interface eth1 --proto udp --dport 445 --jump ACCEPT

iptables -L
service iptables save

얼마전에 "웹진화론 2" 란 책을 추천 받아서 읽어 보았다.
작가(우메다 모치오)가 좌우명처럼 생각하는 말 중에 하나가,, 이 말이다.
직역하면,, "the Paranoid 만 살아남는다.."
the Paranoid 가 그럼 뭐냐??
사전적 의미는 편집병자 이다..
즉 정상적인 사람은 아니다..
뭔가에 집착하고, 지나치게 병적으로 의심하는 사람..
Intel 의 전 회장이자 현 상임고문인 앤드루 그로브가 한 말이다.
나도 그럴 수 있을까?
매사에 놀라운 완벽주의와 집요한 실행력을 가진 자만이 살아남을 수 있는 것일까??
나의 좌우명은 뭐지?


출처 : http://www.intel.com/pressroom/kits/bios/grove/paranoid.htm

Only the Paranoid Survive: Book Preface

Sooner or later, something fundamental in your business world will change.

I'm often credited with the motto, "Only the paranoid survive." I have no idea when I first said this, but the fact remains that, when it comes to business, I believe in the value of paranoia. Business success contains the seeds of its own destruction. The more successful you are, the more people want a chunk of your business and then another chunk and then another until there is nothing left. I believe that the prime responsibility of a manager is to guard constantly against other people's attacks and to inculcate this guardian attitude in the people under his or her management.

The things I tend to be paranoid about vary. I worry about products getting screwed up, and I worry about products getting introduced prematurely. I worry about factories not performing well, and I worry about having too many factories. I worry about hiring the right people, and I worry about morale slacking off.


And, of course, I worry about competitors. I worry about other people figuring out how to do what we do better or cheaper, and displacing us with our customers.

But these worries pale in comparison to how I feel about what I call strategic inflection points.

I'll describe what a strategic inflection point is a bit later in this book. For now, let me just say that a strategic inflection point is a time in the life of a business when its fundamentals are about to change. That change can mean an opportunity to rise to new heights. But it may just as likely signal the beginning of the end.

Strategic inflection points can be caused by technological change but they are more than technological change. They can be caused by competitors but they are more than just competition. They are full-scale changes in the way business is conducted, so that simply adopting new technology or fighting the competition as you used to may be insufficient. They build up force so insidiously that you may have a hard time even putting a finger on what has changed, yet you know that something has. Let's not mince words: A strategic inflection point can be deadly when unattended to. Companies that begin a decline as a result of its changes rarely recover their previous greatness.

But strategic inflection points do not always lead to disaster. When the way business is being conducted changes, it creates opportunities for players who are adept at operating in the new way. This can apply to newcomers or to incumbents, for whom a strategic inflection point may mean an opportunity for a new period of growth.

You can be the subject of a strategic inflection point but you can also be the cause of one. Intel, where I work, has been both. In the mid-eighties, the Japanese memory producers brought upon us an inflection point so overwhelming that it forced us out of memory chips and into the relatively new field of microprocessors. The microprocessor business that we have dedicated ourselves to has since gone on to cause the mother of all inflection points for other companies, bringing very difficult times to the classical mainframe computer industry. Having both been affected by strategic inflection points and having caused them, I can safely say that the former is tougher. I've grown up in a technological industry. Most of my experiences are rooted there. I think in terms of technological concepts and metaphors, and a lot of my examples in this book come from what I know. But strategic inflection points, while often brought about by the workings of technology, are not restricted to technological industries.

The fact that an automated teller machine could be built has changed banking. If interconnected inexpensive computers can be used in medical diagnosis and consulting, it may change medical care. The possibility that all entertainment content can be created, stored, transmitted and displayed in digital form may change the entire media industry. In short, strategic inflection points are about fundamental change in any business, technological or not.

We live in an age in which the pace of technological change is pulsating ever faster, causing waves that spread outward toward all industries. This increased rate of change will have an impact on you, no matter what you do for a living. It will bring new competition from new ways of doing things, from corners that you don't expect.

It doesn't matter where you live. Long distances used to be a moat that both insulated and isolated people from workers on the other side of the world. But every day, technology narrows that moat inch by inch. Every person in the world is on the verge of becoming both a coworker and a competitor to every one of us, much the same as our colleagues down the hall of the same office building are. Technological change is going to reach out and sooner or later change something fundamental in your business world.

Are such developments a constructive or a destructive force? In my view, they are both. And they are inevitable. In technology, whatever can be done will be done. We can't stop these changes. We can't hide from them. Instead, we must focus on getting ready for them. The lessons of dealing with strategic inflection points are similar whether you're dealing with a company or your own career. If you run a business, you must recognize that no amount of formal planning can anticipate such changes. Does that mean you shouldn't plan? Not at all. You need to plan the way a fire department plans: It cannot anticipate where the next fire will be, so it has to shape an energetic and efficient team that is capable of responding to the unanticipated as well as to any ordinary event. Understanding the nature of strategic inflection points and what to do about them will help you safeguard your company's well-being. It is your responsibility to guide your company out of harm's way and to place it in a position where it can prosper in the new order. Nobody else can do this but you. If you are an employee, sooner or later you will be affected by a strategic inflection point. Who knows what your job will look like after cataclysmic change sweeps through your industry and engulfs the company you work for? Who knows if your job will even exist and, frankly, who will care besides you?

Until very recently, if you went to work at an established company, you could assume that your job would last the rest of your working life. But when companies no longer have lifelong careers themselves, how can they provide one for their employees?

As these companies struggle to adapt, the methods of doing business that worked very well for them for decades are becoming history. Companies that have had generations of employees growing up under a no-layoff policy are now dumping 10,000 people onto the street at a crack. The sad news is, nobody owes you a career. Your career is literally your business. You own it as a sole proprietor. You have one employee: yourself. You are in competition with millions of similar businesses: millions of other employees all over the world. You need to accept ownership of your career, your skills and the timing of your moves. It is your responsibility to protect this personal business of yours from harm and to position it to benefit from the changes in the environment. Nobody else can do that for you.

Having been a manager at Intel for many years, I've made myself a student of strategic inflection points. Thinking about them has helped our business survive in an increasingly competitive environment. I'm an engineer and a manager, but I have always had an urge to teach, to share with others what I've figured out for myself. It is that same urge that makes me want to share the lessons I've learned.

This book is not a memoir. I am involved in managing a business and deal daily with customers and partners, and speculate constantly about the intentions of competitors. In writing this book, I sometimes draw on observations I have made through such interactions. But these encounters didn't take place with the notion that they would make it into any public arena. They were business discussions that served a purpose for both Intel and others' businesses, and I have to respect that. So please forgive me if some of these stories are camouflaged in generic descriptions and anonymity. It can't be helped.

What this book is about is the impact of changing rules. It's about finding your way through uncharted territories. Through examples and reflections on my and others' experiences, I hope to raise your awareness of what it's like to go through cataclysmic changes and to provide a framework in which to deal with them.

Copyright © 1996 by Andrew S. Grove. All rights reserved. 

'ETC' 카테고리의 다른 글

ATTO  (0) 2011.06.16
이 또한 지나가리라  (0) 2010.11.25
휴브리스(hubris), 과거의 작은 성공을 잊어라  (0) 2009.11.17
가을인데,, 책읽기에 빠져 보심이...  (0) 2009.09.15
Lenovo T61 Multibooting  (4) 2009.07.22
Rocks 5.2 + RHEL 5.3 (x86_64) 버전
SGE-v62u2_1-1

qrsh 를 실행하면 아래와 같이 에러가 발생합니다.

error: error: ending connection before all data received
error:
error reading job context from "qlogin_starter"

아래 내용을 근거로 Queue 설정에서 아래쪽 두줄을 삭제합니다.

rsh_command                  /usr/bin/ssh
rlogin_command               /usr/bin/ssh


[root@frontend ~]# qconf -sconf
#global:
execd_spool_dir              /opt/gridengine/default/spool
mailer                       /bin/mail
xterm                        /usr/bin/X11/xterm
load_sensor                  none
prolog                       none
epilog                       none
shell_start_mode             posix_compliant
login_shells                 sh,ksh,csh,tcsh
min_uid                      0
min_gid                      0
user_lists                   none
xuser_lists                  none
projects                     none
xprojects                    none
enforce_project              false
enforce_user                 auto
load_report_time             00:00:40
max_unheard                  00:05:00
reschedule_unknown           00:00:00
loglevel                     log_warning
administrator_mail           none
set_token_cmd                none
pag_cmd                      none
token_extend_time            none
shepherd_cmd                 none
qmaster_params               none
execd_params                 none
reporting_params             accounting=true reporting=true \
                             flush_time=00:00:15 joblog=true sharelog=00:00:00
finished_jobs                100
gid_range                    20000-20100
qlogin_command               builtin
qlogin_daemon                builtin
rlogin_command               builtin
rlogin_daemon                builtin
rsh_command                  builtin
rsh_daemon                   builtin
max_aj_instances             2000
max_aj_tasks                 75000
max_u_jobs                   0
max_jobs                     0
max_advance_reservations     0
auto_user_oticket            0
auto_user_fshare             0
auto_user_default_project    none
auto_user_delete_time        86400
delegated_file_staging       false
reprioritize                 0
jsv_url                      none
qrsh_command                 /usr/bin/ssh
rsh_command                  /usr/bin/ssh
rlogin_command               /usr/bin/ssh
jsv_allowed_mod              ac,h,i,e,o,j,M,N,p,w

[root@frontend ~]# qconf -mconf
#rsh_command                  /usr/bin/ssh   ## 삭제
#rlogin_command               /usr/bin/ssh   ## 삭제

얼마전에 "성공 경험의 자기 우상화" 라는 말을 들었습니다.
과거에 이런 방식으로 문제에 접근했을 때 모든 일이 잘 해결되었기 때문에,
현재 나에게 일어나는 문제에 대해서도 동일한 방식으로 접근한다??
너무도 당연한 논리이다..
그런데,, 이런 접근 방식이 모두 성공하지는 않을텐데... 걱정이다.
아래 독재자의 예 처럼...........

 

휴브리스(hubris), 과거의 작은 성공을 잊어라

2009-09-24

[한국경제신문 2009.2.10일자]

 

모든 성공은 최면이요 마약이다. 언제든 반복될 수 있고 어디서든 통할 것만 같다. 그러나 경쟁이 있는 사회에서 그런 일은 잘 안 생긴다.

과거의 성공은 사다리일 뿐이다. 올라오는데 쓴 도구는 더 높은 곳으로 올라가는 데는 쓸모가 없다. 성공이란 사다리를 걷어찰 수 있어야 진정한 CEO의 길에 들어설 수 있다.

회사에서 2월은 의욕이 넘치는 시절이다. 경영진에 막 합류한 신임 임원들의 '고양된' 사기가 곳곳에서 넘친다.

그러나 스포츠에서 쉽게 알 수 있듯 흥분만큼 위험한 것도 없다. 고수를 만나면 여지없이 깨지고,엔돌핀이 너무 높아져 부하들과 코드가 안 맞는다. 더군다나 최악의 경기 상황이 한동안 계속될 수밖에 없는 현실이어서 성과도 기대만큼 나타나지 않는다.

장차 위대한 CEO가 되려면 초임 임원 시절부터 침착해야 한다. 방향을 잘 잡고 있는지 항상 자문해야 한다. 고수의 세계에선 한번의 실수로 승부가 갈린다. 특히 조심해야 할 대목은 과거의 성공 경험을 고집하는 것이다. 부장 시절에 재미를 본 방법을 계속 써먹다간 '부장감'으로 격하되고,목숨이 끝날 수도 있다.

사실 초임 임원들에만 해당되는 것은 아니다. 모든 리더는 과거의 성공 경험을 자신의 밑천으로 안다. 그것이 경험이라고 생각한다.

그러나 이렇게 변화 빠른 시절에 과거의 성공은 그야말로 과거일 뿐이다. 쉽게 모방되고 비결이 알려지는 시대에는 한번 성공한 것이 오히려 독이 되기도 한다.

역사학자 토인비는 역사를 바꾸는데 성공한 창조적 소수가 그 성공으로 인해 교만해져서 남의 말에 귀를 막고 독단적으로 행동하다 판단력을 잃게 되는 것
휴브리스(hubris)라고 불렀다.

원래의 뜻인 오만,자만을 넘어 '성공 경험의 자기 우상화'라고 부를 수 있는 현상이다. 총칼로 권력을 잡은 군인이 계속 그 방식으로 철권통치를 하다 결국 쿠데타로 실각하는 것이 휴브리스의 대표적인 사례다.

달콤한 유혹을 뜻하는 '마시멜로'도 같은 맥락이다. 어린 아이에게 마시멜로를 주고 15분간 안 먹고 참으면 하나 더 주겠다는 스탠퍼드대의 실험에선 3분의 1의 아이들이 참지 못하고 먹어치웠다. 유혹을 견뎌낸 아이들이 결국 크게 성공을 거뒀다.

성공은 이미 과거다. 과거의 경험이 통하지 않는 시대엔 성공도 버려야 할 기억이다. 과거의 내 방식이 더 이상 통하지 않을 것이라고 믿을 때 더욱 비장해진다. 마시멜로의 유혹을 이겨내며 휴브리스에 얽매이지 않을 때라야 당신의 경영에도 날개가 달릴 것이다.

권영설 한경 아카데미 원장 yskwon@hankyung.com
한경 가치혁신연구소 value.hankyung.com

'ETC' 카테고리의 다른 글

ATTO  (0) 2011.06.16
이 또한 지나가리라  (0) 2010.11.25
Only the Paranoid Survive - Andrew S. Grove  (0) 2009.11.17
가을인데,, 책읽기에 빠져 보심이...  (0) 2009.09.15
Lenovo T61 Multibooting  (4) 2009.07.22

인터넷 검색을 하다가,,
저와 똑같은 기종인 T61p에서 8GB RAM을 사용하는 분을 발견했습니다.
역시,, 64bit OS를 사용하니까,,,
노트북에서도 8GB를 쓰는군요..
대단하십니다.

http://blogs.technet.com/keithcombs/archive/2008/06/04/my-lenovo-thinkpad-t61p-now-has-8gb-of-ram.aspx

'Windows' 카테고리의 다른 글

GVIM mswin.vim 환경설정  (0) 2010.12.13
윈도우즈 무료 백업 솔루션 ZConverter Windows Backup  (0) 2009.09.15
서류 가방 에러  (0) 2009.09.01
tasklist 로 실행 프로세스 확인  (0) 2009.07.01
Windows Command Line Scripts  (0) 2009.06.21

얼마전에,, 책상에서 뒹굴던 책 한권이 눈에 들어왔다

"부코스키가 간다"

뭔가 읽어야 한다는 생각에 무작정 읽기 시작했는데, 나름 인상에 남는다..
88만원 세대라고 일컬어 지는 요즘 세대들.. 백수라는 직업,, 안타까움,,
조금 일찍 태어난 내가 어찌나 다행스러운지...

암튼 이 책을 계기로,, 책꽂이에 꽂혀있는 책들을 한권 두권 읽기 시작했다.
다음으로 손에 잡힌건,,,

"엄마를 부탁해" - 신경숙 지음.

책 읽은지 한달이 지났지만,, 아직도 그 내용을 생각하면 눈물이 핑 돈다.
시골에서 자란 나로서는 그 내용이 누구보다도 마음에 와 닿았다.

암튼,, 요즘엔 출퇴근 시간에 책 읽는 습관을 들였다
약 40분 정도 지하철을 타는데,, 그동안 아무 생각 없이 지하철 타고 다니던 시간이 너무나 아깝다.

다음으로,, 장외인간(이외수)를 읽었고,, 요즘엔 역사소설이 빠져있다.
(춘천에서 고등학교를 다닐 때, 이외수 선생님이 학교에 오셔서 특강을 한 적이 있다.
 장외인간은 춘천을 배경으로 하고 있어서,, 글 내용이 눈앞에 그려진다..
 다만,, 이외수 선생님의 독특한 발상이 잘 적응이 안된다..)

상도, 해신은 드라마로도 제작 방영되어 가끔 보기도 했는데,,
원작자가 최인호 선생님인것은 "잃어버린 왕국"을 보고 있는 요즘에야 알게되었다.
정말 그동안 책을 너무 멀리하지 않았나,, 헐~~~~
3권을 읽는 중인데,, 황산벌에서 김유신과 계백이 만나서 전투를 벌이는 중이다..
퇴근 시간에 빨리 읽어 봐야 하는데,,, ㅎㅎㅎ

'ETC' 카테고리의 다른 글

ATTO  (0) 2011.06.16
이 또한 지나가리라  (0) 2010.11.25
Only the Paranoid Survive - Andrew S. Grove  (0) 2009.11.17
휴브리스(hubris), 과거의 작은 성공을 잊어라  (0) 2009.11.17
Lenovo T61 Multibooting  (4) 2009.07.22

ZConverter 에서 무료 백업 솔루션이 발표되었습니다.
2009년 9월 15일 부터 무료로 다운로드 받을 수 있다고 하는 군요..

http://www.zconverter.co.kr/ 에서 회원가입하면 다운로드 받을 수 있습니다.

회원가입할 때 추천인 아이디에 jklee 를 입력하시면,
Essential License 를 보내줍니다.

회원가입 바로가기 : http://www.zconverter.co.kr/bbs/register.php
추천인 아이디에 반드시 jklee 를 입력하세요..

'Windows' 카테고리의 다른 글

GVIM mswin.vim 환경설정  (0) 2010.12.13
My Lenovo ThinkPad T61p now has 8GB of RAM  (0) 2009.09.15
서류 가방 에러  (0) 2009.09.01
tasklist 로 실행 프로세스 확인  (0) 2009.07.01
Windows Command Line Scripts  (0) 2009.06.21
혹시 윈도우의 기본 기능 중 서류 가방이라고 들어보셨나요?

VISTA 도움말을 찾아보니,,아래와 같이 설명되어 있더군요...

서류 가방을 사용하여 동기화

서류 가방을 사용하면 컴퓨터가 같은 네트워크에 없는 경우에도 서로 다른 두 컴퓨터 간에 파일을 동기화할 수 있습니다. 컴퓨터가 같은 네트워크에 없을 경우 이동식 미디어를 사용하여 컴퓨터 간에 파일을 복사하고, 다른 컴퓨터를 사용하여 파일에서 작업을 수행한 다음, 서류 가방을 사용하여 변경 내용을 원본 컴퓨터로 다시 동기화할 수 있습니다.

저는 서류 가방을 USB 디스크 폴더와 로컬 디스크 폴더의 동기화를 위해 사용합니다.

예를 들어,, 로컬 디스크에 D:\DOCUMENTS 라는 폴더를 만들고,,
모든 문서작업은 이 폴더 아래에 저장해 두었습니다.
백업을 위해 D:\DOCUMENTS 폴더를 외장하드 F:\DOCUMENTS 에 복사를 하죠..
그런데,, 다른 컴퓨터에서 외장하드안에 있는 문서 중 일부를 수정했습니다.
또,, D:\DOCUMENTS 에도 일부 파일이 수정되었구요..
양쪽에서 수정 작업을 하면,, 어떤 파일이 최신 파일인지 찾아내기 힘들죠?

서류가방을 쓰면 간단하게 동기화 할 수 있습니다.
외장하드 F:\DOCUMENTS 라는 폴더를 만들고,,
D: 에는 DOCUMENTS 라는 서류가방을 만듭니다.
그리고 나서,, F:\DOCUMENTS 폴더 아래에 있는 모든 내용을
Drag & Drop으로 서류가방 아래에 가져다 둡니다.
그러면,, 윈도우가 알라서 변경된 내용을 찾아내서 업데이트 합니다.
양쪽에서 변경된 내용이 있을 경우에는 어떻게 할 지 물어봅니다.
상당히 편리하죠.. ㅎㅎㅎ

그런데,, 어제부터,, 이상한 에러가 발생했습니다.


잘 쓰던 놈이 에러를 뱉어 내면서, 튕겨나가는데,, 머리 아픕니다.
이 넘을 해결하던가,,
아니면,, 새로운 방법을 찾아야죠..

요즘엔 robocopy 라는 명령어가 유용해 보이기도 하는데,,, 아직은 서류 가방이 더 편리...


Windows HPC Server 2008 R2 CTP 버전이 지난 5월에 올라왔습니다.

크게 달라진 점은,,

1. Windows Server 2008 R2 기반
2. Job Scheduling, runtimes, and management 관련 확장성과 관리성 향상

http://www.microsoft.com/hpc/en/us/ctp-download.aspx

링크에 접속하시면 CTP 버전을 다운로드 받으실 수 있습니다.

Job Scheduler

The job scheduler scales to a larger number of nodes and processing cores, while boosting overall cluster utilization through more sophisticated scheduling policies. Some examples present in the CTP include:

  • Scheduling at Scale and improved task throughput

  • Ability to monitor Job Progress with greater granularity

  • Ability to set up Pre and Post Tasks

Runtimes

Runtime support has now been expanded to include more robust programming models as well improved support for jobs based on the SOA programming model. Examples present in the CTP include:

  • Fire and Recollect Programming Model

  • Fire and Recollect Programming Model WCF Broker Re-architecture to permit job throttling etc.

  • Greater interoperability with Java based jobs

Management

With Windows HPC Server 2008 R2, you get a scalable platform both from a deployment and system management perspective as well as extensibility of the reporting functions to ensure more productive system management. Some examples included in the CTP are:

  • Deployment at large node counts

  • Efficient Management at large node counts

  • Customizable Heat Map for better system wide monitoring

  • Extensible reporting to tailor HPC infrastructure reports for specific requirements


HPC Pack 2008 Service Pack 1이 나왔네요..
지난 7월 7일 올라왔구요..
아래 링크에서 다운로드 받을 수 있습니다.
Windows Update를 통해서도 자동 업데이트 됩니다.

http://www.microsoft.com/downloads/details.aspx?FamilyID=a78520f1-de82-4442-b273-2ff7327b9140&displaylang=en

Quick Details
Version: 2.1.1703
Date Published: 7/7/2009
Language: English
Download Size: 15.9 MB - 32.2 MB*

Files in This Download

The links in this section correspond to separate files available in this download. Download the files most appropriate for you.
File Name: File Size

HPC2008_SP1-x64.exe

16.3 MB

HPC2008_SP1-x86.exe

15.9 MB

제가 사용하는 노트북은 Lenovo T61p 입니다.
Wide Screen 이라서,, 무게가 상당히 나가는데,, 매일 들고 다닌답니다. 헉헉..
이것 저것 테스트할 일이 많아서,, 가상머신을 많이 사용하는데,,
요즘엔 가상머신 종류도 다양해서,, 어려움이 많았습니다.
해결 방법으로 찾은 것이 2nd HDD Bay... ^^

CD-ROM Drive








T-Series 오른쪽 옆면에 보면 CD-ROM Bay가 있습니다. 잡아 빼면 빠지게 되어 있죠..
CD-ROM을 제거하고, 2nd HDD Bay를 장착하면 추가 하드디스크를 사용할 수 있습니다.


Serial Hard Drive Bay Adapter




종류가 여러가지 있겠지만,, 요즘엔 SATA가 대세라,,
Serial Hard Drive Bay Adapter (FRU/N 26R9246, P/N 26R9247)을 구매했습니다.
오래전에 구매해서 가격은 기억이 안나네요..
처음엔 삼성 250GB 5400rpm 제품을 구매해서,, 가상머신 전용으로 사용했습니다. VMware workstation 버전을 사용하고 있는데,, 250GB 만으로도 부족하더군요..
게다가,, Hyper-V, Xen 등등도 테스트 해야 하고,,


SATA HDD

지름신께서 왕림하사,, HDD를 추가 구매했습니다.
SAMSUNG 250 GB 5400 rpm
SAMSUNG 320 GB 5400 rpm
WesternDigital 320 GB 7200 rpm

이넘들을 어떻게 사용하냐구요??


1. 100GB (7200rpm) - Hitachi (내장 HDD)
- System Recovery
- (C:) Windows VISTA
- (D:) DATA

lenovo HDD
2. 320GB (7200rpm) - WesternDigital
- Windows VISTA
- Windows 7
- VMs

3. 320GB (5400rpm) - SAMSUNG
- Windows 7
- Windows Server 2008 EE
- CentOS
- VMs

4. 250GB (5400rpm) - SAMSUNG
- VMs

자,, 멀티 부팅 환경을 한번 보실까요??
먼저 노트북 부팅 후에 F12 키를 누르면, 부트 디바이스를 선택하는 화면이 나옵니다.
Boot Menu


2nd HDD Bay 에 장착된 HDD가 보이죠?? 이넘을 선택하면 두번째 디스크로 부팅할 수 있죠..
Grub Menu


CentOS가 깔려있는 디스크의 경우 GRUB을 2nd HDD 에 심어줘야 합니다.
설치할 때 조심해야 할 점은 내장 디스크 MBR을 건드리지 않는 것이죠..
가능하면, 2nd HDD에 OS를 설치할 때 내장 디스크를 빼고 진행하는 것도 좋은 방법입니다.
위에 BIOS 메뉴에서,,

4: ATA HDD0 : 내장 HDD
5: ATA HDD1 : 2nd HDD

라고 나오는데,, CentOS에서는 순서가 바뀝니다.
아래 그림처럼,, 2nd HDD 를 HD0로 인식하죠..

Windows 2008 EE


내장 디스크는 HD1으로 인식됩니다. 그래서,, 내장 디스크의 VISTA로 부팅하기 위해서
GRUB에서는 아래와 같이 설정합니다.
Windows VISTA

내장디스크의 첫번째 파티션은 ThinkVantage 에서 잡고 있죠..
ThinkVantage 로 부팅하는 메뉴도 추가해 두었죠.. ^^
ThinkVantage



GRUB을 사용하면 부트 메뉴 편집하기가 쉬운데,,
Windows 7, VISTA, Windows 2008 등을 사용하면,,
BCDEDIT를 사용해야 합니다.
처음에, 이넘때문에 고생 많이 했습니다.
암튼 지금은 노트북으로 다양한 OS를 즐기고(?) 있습니다.

'ETC' 카테고리의 다른 글

ATTO  (0) 2011.06.16
이 또한 지나가리라  (0) 2010.11.25
Only the Paranoid Survive - Andrew S. Grove  (0) 2009.11.17
휴브리스(hubris), 과거의 작은 성공을 잊어라  (0) 2009.11.17
가을인데,, 책읽기에 빠져 보심이...  (0) 2009.09.15

신규 서버를 도입한 이후에 성능이 예전만 못하다!!
돈들여 구입했는데,, 예전만 못하다면,, 실망이 이만 저만이 아니죠..
분명히 최근에 나온 신규 서버이고,,
Intel에서 CPU의 성능이 월등히 좋아졌다고 하는데,,
2년전에 도입한 서버보다도 성능이 떨어지다니!!!

여기서 잠깐,,,
성능비교 환경을 잠깐 살펴보죠..

1. CPU
   OLD : Intel 3000-series "Conroe" 3070 (2.66GHz Dual-core, 4MB L2 Cache, 1066MHz FSB)
   NEW : Intel 5500-series "Nehalem-EP" (Gainestown) E5504 (2.0GHz Quad-core, 4MB L3 Cache)

2. Application
   User Developed MPI Program. USE 4 cores

아~~~ 눈물이 앞을 가립니다.
Dual-core <--> Quad-core
"Conroe" <--> "Nehalem-EP"
어느것 하나 떨어지는 것이 없는데,, 왜 성능이 떨어지는 것입니까????

이것이 바로 성능 비교의 오류입니다.
사용하는 application은 MPI 프로그램으로 CPU를 많이 사용합니다.
비교에 사용된 core 개수는 OLD / NEW 모두 4 core를 사용했습니다.
4 core를 사용할 경우 초당 부동소수점 연산 (FLOPS)를 이론치는,,

Rpeak (OLD) = 2.66 * 4 cores * 4 = 42.56
Rpeak (NEW) = 2.0 * 4 cores * 4 = 32

같은 core를 사용할 경우 성능을 좌우하는 것은 CPU Clock 입니다.
Clock 이 낮으므로,, 신규 서버의 성능이 당연히 떨어질 수 밖에 없습니다.
당연한 것 때문에,, 고민하는 분이,,, 앞으로는 안계시길...
tasklist


리눅스에서는 ps 명령으로 프로세스 리스트를 볼 수 있는데,,
윈도우즈에서도 cmd 창에서 실행할 수 있는 명령어가 있을까?
tasklist 를 사용하면 됩니다.

/FI 필터 옵션을 통해 원하는 프로세스만 잡아낼 수도 있습니다.

tasklist /fi "imagename eq outlook.exe" /m

'Windows' 카테고리의 다른 글

윈도우즈 무료 백업 솔루션 ZConverter Windows Backup  (0) 2009.09.15
서류 가방 에러  (0) 2009.09.01
Windows Command Line Scripts  (0) 2009.06.21
Remote Desktop 공유하기  (0) 2009.06.21
[WS2008] 원격 데스크톱의 변화  (0) 2009.06.21

+ Recent posts